emu/common: add function to read utf16 file

EKA2L1 · Feb 28, 2020 · b9a8204 · b9a8204
1 parent 934466b
commit b9a8204
Show file tree

Hide file tree

Showing 2 changed files with 28 additions and 7 deletions.
diff --git a/src/emu/common/include/common/fileutils.h b/src/emu/common/include/common/fileutils.h
@@ -52,6 +52,10 @@ namespace eka2l1::common {
     */
     bool remove(const std::string &path);
 
+    /* !\brief Get the content of a UTF-16 LE encoded file as UTF8 string.
+    */
+    std::string get_file_content_u16(const std::string &path);
+
     bool move_file(const std::string &path, const std::string &new_path);
 
     /**

diff --git a/src/emu/common/src/fileutils.cpp b/src/emu/common/src/fileutils.cpp
@@ -240,7 +240,7 @@ namespace eka2l1::common {
 
         do {
             cycles_to_next_entry();
-            d = reinterpret_cast<struct dirent*>(find_data);
+            d = reinterpret_cast<struct dirent *>(find_data);
         } while (d && (strncmp(d->d_name, ".", 1) == 0 || strncmp(d->d_name, "..", 2) == 0));
 #endif
 
@@ -306,14 +306,32 @@ namespace eka2l1::common {
 #endif
     }
 
+    std::string get_file_content_u16(const std::string &path) {
+        std::ifstream fin(path, std::ios::binary);
+        fin.seekg(0, std::ios::end);
+        size_t size = static_cast<size_t>(fin.tellg());
+        if (size <= 2) return "";
+
+        //skip BOM
+        fin.seekg(2, std::ios::beg);
+        size -= 2;
+
+        std::u16string u16((size / 2) + 1, '\0');
+        fin.read(reinterpret_cast<char*>(&u16[0]), size);
+
+        std::string utf8 = std::wstring_convert<std::codecvt_utf8_utf16<char16_t>, char16_t>{}.to_bytes(u16);
+
+        return utf8;
+    }
+
     bool move_file(const std::string &path, const std::string &new_path) {
 #if EKA2L1_PLATFORM(WIN32)
         return MoveFileA(path.c_str(), new_path.c_str());
 #else
         return (rename(path.c_str(), new_path.c_str()) == 0);
 #endif
     }
-    
+
     bool copy_file(const std::string &target_file, const std::string &dest, const bool overwrite_if_dest_exists) {
 #if EKA2L1_PLATFORM(WIN32)
         return CopyFile(target_file.c_str(), dest.c_str(), !overwrite_if_dest_exists);
@@ -325,7 +343,7 @@ namespace eka2l1::common {
         if (eka2l1::exists(dest) && !overwrite_if_dest_exists) {
             return false;
         }
-            
+
         std::ifstream src(target_file, std::ios::binary);
         std::ofstream dst(dest, std::ios::binary | std::ios::trunc);
 
@@ -344,13 +362,12 @@ namespace eka2l1::common {
         if (GetFileAttributesExW(reinterpret_cast<const LPCWSTR>(path.c_str()), GetFileExInfoStandard, &attrib_data) == false) {
             return 0xFFFFFFFFFFFFFFFF;
         }
-                
+
         FILETIME last_modify_time = attrib_data.ftLastWriteTime;
 
         // 100 nanoseconds = 0.1 microseconds
         return convert_microsecs_win32_1601_epoch_to_1ad(
-            static_cast<std::uint64_t>(last_modify_time.dwLowDateTime) | 
-            (static_cast<std::uint64_t>(last_modify_time.dwHighDateTime) << 32));
+            static_cast<std::uint64_t>(last_modify_time.dwLowDateTime) | (static_cast<std::uint64_t>(last_modify_time.dwHighDateTime) << 32));
 #else
         const std::string name_utf8 = common::ucs2_to_utf8(path);
         struct stat st;
@@ -363,7 +380,7 @@ namespace eka2l1::common {
         return convert_microsecs_epoch_to_1ad(static_cast<std::uint64_t>(st.st_mtime));
 #endif
     }
-    
+
     bool is_system_case_insensitive() {
 #if EKA2L1_PLATFORM(WIN32)
         return true;